Abstract
The difficulties and instabilities accompanying the inversion of radiance data to infer temperature structure are closely related to the high degree of interdependence existing among these nominally independent measurements.
The radiance measurements discussed in the accompanying papers are shown to be interdependent to a marked degree. It is shown that one of the measurements can be predicted from the others with an accuracy which is only a little worse than the experimental accuracy.
The application of this kind of analysis to determine optimum choices of measurements and the information content thereof is outlined.
